1,720,963 research outputs found
vnep-approx/evaluation-ifip-networking-2018: Code and Data for Paper: "Virtual Network Embedding Approximations: Leveraging Randomized Rounding"
No full text<p>This release of our evaluation of approximation algorithms for the Virtual Network Embedding Approximation Algorithm is final: it contains all necessary sources to reproduce the results of our paper presented at IFIP Networking 2018. Besides the reduced data used for plotting, the release additionally provides:</p>
<ul>
<li>the generated scenarios (ifip_networking_evaluation_scenarios.pickle.tar.gz, 97 MB, 522 MB extracted)</li>
<li>the MIP baseline solutions (ifip_networking_evaluation_solutions_MIP_baseline.pickle.tar.gz, 132 MB, 659 MB extracted) </li>
<li>the randomized rounding solutions (ifip_networking_evaluation_solutions_rand_round.pickle.tar.gz, 171 MB, 864 MB extracted) </li>
</ul>
Is There a Case for Parallel Connections with Modern Web Protocols?
Full text linkModern web protocols like HTTP/2 and QUIC aim to make the web faster by addressing well-known problems of HTTP/1.1 running on top of TCP. Both HTTP/2 and QUIC are specified to run on a single connection, in contrast to the usage of multiple TCP connections in HTTP/1.1. Reducing the number of open connections brings a positive impact on the network infrastructure, besides improving fairness among applications. However, the usage of a single connection may result in poor application performance in common adverse scenarios, such as under high packet losses. In this paper we first investigate these scenarios, confirming that the use of a single connection sometimes impairs application performance. We then propose a practical solution (here called H2-Parallel) that implements multiple TCP connection mechanism for HTTP/2 in Chromium browser. We compare H2-Parallel with HTTP/1.1 over TCP, QUIC over UDP, as well as HTTP/2 over Multipath TCP, which creates parallel connections at the transport layer opaque to the application layer. Experiments with popular live websites as well as controlled emulations show that H2-Parallel is simple and effective. By opening only two connections to load a page with H2-Parallel, the page load time can be reduced substantially in adverse network conditions
SRv6Pipes: enabling in-network bytestream functions
No full textIPv6 Segment Routing is a recent IPv6 extension that is generating a lot of interest among researchers and in industry. Thanks to IPv6 SR, network operators can better control the paths followed by packets inside their networks. This provides enhanced traffic engineering capabilities and is key to support Service Function Chaining (SFC). With SFC, an end-to-end service is the composition of a series of in-network services. Simple services such as NAT, accounting or stateless firewalls can be implemented on a per-packet basis. However, more interesting services like transparent proxies, transparent compression or encryption, transcoding, etc. require functions that operate on the bytestream. In this paper, we extend the IPv6 implementation of Segment Routing in the Linux kernel to enable network functions that operate on the bytestream and not on a per-packet basis. Our SRv6Pipes enable network architects to design end-to-end services as a series of in-network functions. We evaluate the performance of our implementation with different microbenchmarks
Tuning Multipath TCP for Interactive Applications on Smartphones
No full textMultipath TCP enables smartphones to simultaneously use both WiFi and LTE to exchange data over a single connection. This provides bandwidth aggregation and more importantly reduces the handover delay when switching from one network to another. This is very important for delay sensitive applications such as the growing voice activated apps. On smartphones, user experience is always a compromise between network performance and energy consumption. However, the Multipath TCP implementation in the Linux kernel was mainly tuned for bandwidth aggregation and often wakes up the cellular interface by creating a path without sending data on it. In this paper, we propose, implement and evaluate MultiMob, a solution providing fast handover with low cellular usage for interactive applications. MultiMob relies on three principles. First, it delays the utilization of the LTE network. Second, it allows the mobile to inform the server of its currently preferred wireless network. Third, MultiMob extends the Multipath TCP handshake to enable immediate retransmissions to speedup handover. We implement MultiMob on Android 6 smartphones and evaluate its benefits by using both microbenchmarks and in the field measurements. Our results show that MultiMob provides similar latency as the standard Linux implementation while significantly lowering the cellular usage
Decentralized Scheduling for Offloading of Periodic Tasks in Mobile Edge Computing
No full textMotivated by various surveillance applications, we consider wireless devices that periodically generate computationally intensive tasks. The devices aim at maximizing their performance by choosing when to perform the computations and whether or not to offload their computations to a cloud resource via one of multiple wireless access points. We propose a game theoretic model of the problem, give insight into the structure of equilibrium allocations and provide an efficient algorithm for computing pure strategy Nash equilibria. Extensive simulation results show that the performance in equilibrium is significantly better than in a system without coordination of the timing of the tasks’ execution, and the proposed algorithm has an average computational complexity that is linear in the number of devices.QC 20180518QC 20200108</p
Going Beyond Counting First Authors in Author Co-citation Analysis
Full text linkThe present study examines one of the fundamental aspects of author co-citation analysis (ACA) - the way co-citation
counts are defined. Co-citation counting provides the data on which all subsequent statistical analyses and mappings
are based, and we compare ACA results based on two different types of co-citation counting - the traditional type that
only counts the first one among a cited work's authors on the one hand and a non-traditional type that takes into
account the first 5 authors of a cited work on the other hand. Results indicate that the picture produced through this non-traditional author co-citation counting contains more coherent author groups and is therefore considerably clearer. However, this picture represents fewer specialties in the research field being studied than that produced through the traditional first-author co-citation counting when the same number of top-ranked authors is selected and analyzed. Reasons for these effects are discussed
Variations on the Author
Full text link“Variations on the Author” discusses two of Eduardo Coutinho’s recent films (Um Dia na Vida, from 2010, and Últimas Conversas, posthumously released in 2015) and their contribution to the general question of documentary authorship. The director’s filmography is characterized by a consistent yet self-effacing form of authorial self-inscription: Coutinho often features as an interviewer that rather than express opinions propels discourses; an interviewer that is good at listening. This mode of self-inscription characterizes him as an author who is not expressive but who is nonetheless markedly present on the screen. In Um Dia na Vida, however, Coutinho is completely absent form the image, while Últimas Conversas, on the contrary, includes a confessional prologue that moves the director from the margins to the center of his films. This article examines the ways in which these works stand out in the filmography of a director who offers new insights into the notion of cinematic authorship
- …
